One day i was working and this tweet got pushed to my phone:
and as always i do have an opinion no it. And the internet needed to know about it, so here it is:
I think that one of the main reasons informal tracking fails is that it’s purpose is not to improve (or at least the stated purpose is to improve in the long term which for all intents and purposes is kind of doomed) but to make you feel good: Did i do my job today ?.
By design this will make you remember only the things which make you feel good and conveniently forget things that didn’t go well. Which actually rubs you of one of the few good lever one can use to improve itself: the knowledge of what went wrong and the ability to dig into why they went wrong.
Formal tracking is better in this respect because it allows you to mentally separate the collection process from the pain of analyzing the data. That way if and when you are ready to look critically at what went didn’t went well you do have the data available and you can dig into it, find out what might not be going well and improve yourself.
Today i found out about an interesting project: “Text and Discourse Understanding: The DISCERN System“. It’s a subsymbolic neural network model of script-based story understanding.
While working on the sentiment analysis tool for the romanian language i always wondered if something like this can work. And it turns out that i can and that it does work pretty well. It also displays human like failures (like schizophrenia for example) which is always a good sign.
And the architecture of the system seems on the right track with what how my intuition suggested which is always good from a self-esteem point of view .
Don’t forget to follow my tumble feed (it’ full of interesting finds that i can’t find the time to properly post them in here)
I have always been interested in AI. I really think that at some point we will be able to build something that will be not that different than what we are currently calling “human mind”.
AI as a term is really oversold at this moment but the field seems to have advanced nicely (and some of the techniques developed have turned out to be really useful in meantime).
So i decided to roll into the “Intro to AI” Online Stanford course. Look interesting enough to worth the effort .
Also if you want to learn about Robotics this should be helpful: artificial intelligence | introduction to robotics.
I noticed something (not really a breakthrough but stay with me). I have two kind of posts:
- the long and convoluted ones (hopefully the product of some thinking)
- the short and off topic ones (made to share some shiny thing i found)
I also have the habit (no idea if good or bad yet) of sharing links to interesting pages on twitter/facebook from time to time (with the same purpose as the second type of post).
In the end I decided to split the things: the long and hopefully thought out posts will remain here and the short and shiny ones are going on a platform much more adapted to it (tumblr.com).
So without further delay i present you my tumblr blog: http://tumble.redeul.ro. It’s fun, it’s shiny, it has some pretense of class and it’s really good for you. Go, read and subscribe to it’s feed: http://tumble.redeul.ro/rss.
Live long and prosper (preferably together).